CloudZero vs ZopNight: Unit Cost Insight or Action on Waste
CloudZero answers 'what does each customer or feature cost us?'. ZopNight answers 'what should be switched off or resized today?'.
What CloudZero does
CloudZero is a cost intelligence platform built around allocation. The most common misreading of it is that you need perfect tagging first. Its own docs say otherwise: “You do not need clean resource tags to get started. Tags are optional enrichment, not a prerequisite” (Organize your costs). Allocation rules are written in CostFormation, which lets you “write rules as code, version control your definitions”, and there is a visual editor alongside it.
It ingests a wide range of spend. The homepage (cloudzero.com) names AWS, GCP, Azure, Kubernetes, Snowflake, Datadog and MongoDB. The docs index adds Oracle, Anthropic and OpenAI, and you can bring any other source through AnyCost, by REST API or S3 (docs index). Optimize “finds savings across your cloud environment and tracks realized impact” (Optimize docs).
Pricing is “a single subscription with all capabilities included”, quoted on request (cloudzero.com/pricing). We found no resource scheduling or automated remediation on the pages we read.
Two different jobs
CloudZero is strongest when the question is commercial. What does serving one customer cost? Which product feature drives AI spend? Is gross margin moving the right way? Its allocation-as-code model is designed so those answers survive messy tagging.
ZopNight overlaps on attribution, but its centre of gravity is operational. It builds showback by team and by tag. It also offers smart tags: virtual tags derived from a resource’s own fields (provider, region, type, instance type, name) under a tagging policy. Once accepted, a smart tag counts toward cost attribution, and you can write it back to the cloud on demand. It has a unit economics overlay too, where you supply a business denominator (orders, active users, API requests) by push API, CSV or pull API. The difference is what happens next. ZopNight schedules idle resources, runs remediation workflows, and records every state change in an audit trail.
| Capability | ZopNight | CloudZero |
|---|---|---|
| Tagging prerequisite | None; smart tags fill gaps from resource fields | None; tags are optional enrichment |
| Allocation model | Team and tag showback, cost-flow Sankey | CostFormation rules as code plus visual editor |
| Non-cloud cost sources | Databricks, Snowflake, AI gateway spend | AnyCost custom sources, Datadog, MongoDB, AI providers |
| Start/stop scheduling | Yes, per resource or group | Not found in docs we read |
| Pricing | See zop.dev/pricing | Single subscription, custom quote |
Choose CloudZero if…
- Your leadership asks for cost per customer, per product or per feature, and allocation fidelity matters more than automation.
- You want allocation logic version-controlled like code.
- A large share of your spend is in SaaS or AI vendors you need to fold into the same model.
Choose ZopNight if…
- You already know where the waste is, and the bottleneck is getting it switched off. Schedules, resource groups and time-boxed overrides turn that into a routine.
- You want recommendations across idle, rightsizing, schedule, orphan, discount, security and reliability categories. Each comes with evidence and a remediation path, including an admin approval gate for destructive steps.
- You want cost anomalies flagged with a root cause, such as an instance resize, a new resource, an expired reservation or a failed schedule, rather than a bare alert.
Questions we get a lot.
If yours isn't here, email us and we'll answer directly.
Does CloudZero require tagging?
No. CloudZero's documentation, checked on 24 September 2026, says clean tags are not needed to get started and that tags are optional enrichment.
Can ZopNight do cost per customer like CloudZero?
Partly. ZopNight's unit economics overlay divides infrastructure cost by a business metric you supply, overall or per team or tag. It is not built for CloudZero-style per-customer allocation modelling, so if that is your main need, CloudZero is the stronger fit.
What does ZopNight do with anomalies?
It evaluates anomalies daily across the org, cloud account, resource type, resource group, resource and team levels, attaches a root-cause analysis to each one, and can notify Slack, Teams, Google Chat or a webhook.